Contemporary society is built on a rapidly expanding number of complex networks. Think of logistics, energy, information and social networks, to name just a few. Each network consists of large numbers of interacting nodes or agents that evolve under the combined forces of their own node-specific dynamics and the influence of their network neighbours. As a consequence, the behaviour of individual agents becomes hard to predict and control. The research effort is focused on trying to characterize global network parameters or performance in terms of the underlying dynamics and topology at the node-level.
In Intelligent Systems, we tackle these issues by drawing methodologies from multi-agent modelling, data analytics, dynamical systems and machine learning. Indeed, although it is frequently possible to explicitly model behaviour at the level of individual nodes, it often requires detailed and extensive simulations to gain insight about the resulting collective network behaviour. Such simulations generate extensive data sets that need to be mined using data analytics and machine learning to uncover interesting patterns.
This research provides a better understanding of how details at the level of individual nodes impact on overall network behaviour. This can then be translated into better engineering decisions to assist in the design and control of such networks. Our research therefore finds applications in network-related problems in smart industry and the optimization of critical infrastructures. It also addresses questions that feature prominently in the Dutch national research and innovation agenda (NWA) route, Resilient and Meaningful Societies (Veerkrachtige en Zinvolle Samenlevingen).